Re: [PATCH 01/12] Miscellaneous functions and defines needed by AppArmor,including the base path resolution routines.

[John Johansen]

John Johansen wrote:
> John Johansen wrote:
>> Al Viro wrote:
>>> On Fri, Feb 19, 2010 at 01:36:17AM -0800, john.johansen@xxxxxxxxxxxxx wrote:
>>>
>>>> +static int d_namespace_path(struct path *path, char *buf, int buflen,
>>>> +			    char **name, int flags)
>>>> +{
>>>> +	struct path root, tmp, ns_root = { };
>>>> +	char *res;
>>>> +	int deleted, connected;
>>>> +	int error = 0;
>>>> +
>>>> +	read_lock(&current->fs->lock);
>>>> +	root = current->fs->root;
>>>> +	/* released below */
>>>> +	path_get(&root);
>>>> +	read_unlock(&current->fs->lock);
>>>> +
>>>> +	spin_lock(&vfsmount_lock);
>>>> +	if (root.mnt && root.mnt->mnt_ns)
>>>> +		/* released below */
>>>> +		ns_root.mnt = mntget(root.mnt->mnt_ns->root);
>>>> +	if (ns_root.mnt)
>>>> +		/* released below */
>>>> +		ns_root.dentry = dget(ns_root.mnt->mnt_root);
>>>> +	spin_unlock(&vfsmount_lock);
>>> Junk.  You might as well leave ns_root {NULL, NULL} instead of that crap.
>>>
>> Right, though the ns_root.mnt and ns_root.dentry are needed below to detect
>> disconnected paths.
>>
> Well no, actually I don't need it.  After looking at this again,
> If struct path tmp is unmodified we hit our expected root.

Sigh so much for "brillant" ideas after a few hours of sleep.  If tmp is modified
we only know that we made it to global_root:   We still need to distinguish lazy
unmount, ideally we would do this in __d_path instead of here.  I'll resurrect our
old patches doing just that and post them out for comment.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/